|
(11) | EP 0 632 584 A1 |
| (12) | EUROPEAN PATENT APPLICATION |
|
|
|
|
|||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||
| (54) | Power amplifier for a radio frequency signal |
| (57) An amplifier is provided comprising: power selection means for selecting a nominal
output power level from a plurality of discrete levels power control means for controlling
output power in response to said power selection means input means for indicating
measured output power; and storage means responsive to the input means for storing,
for each of said nominal output power levels, information in response to the measured
output power, for future adjustment of each nominal output power level when selected. |
[0001] This invention provides a power amplifier for amplifying a radio frequency signal, for example a pulsed power amplifier responsive to a control pulse. The amplifier is particularly useful for digital mobile cellular radio transmitters for use on the Pan-European GSM cellular network.
[0002] In the field of power control circuits, EP-A-0261967 describes a control loop for a power amplifier in which a digital value is output from a controller and converted to analog form for control of the amplifier when input power falls below a minimum threshold. Power at the beginning of a new burst is the same as at the end of a previous burst.
[0003] In a burst modulated power amplifier, the transmitter must observe a time domain template upon turn-on and turn-off, as well as a frequency domain template. In the past, the shape of the power characteristic as it rises at the start of a burst and falls at the end has been controlled by means of shaping circuits consisting of resistors and analog switches. Such circuits can be bulky and unreliable and have limited accuracy.
[0004] As well as the above power/time characteristic, the output power of a GSM mobile radio transmitter must be adjustable in sixteen steps from +43 dbm to +13 dbm. Many tolerance factors within the amplifier will affect the ultimate output power. Manual adjustment means can be provided for pre-setting the output power levels before the equipment leaves the factory, however separate adjustment means within the equipment for each of the sixteen power levels would be bulky, and their adjustment would be time consuming.
[0005] It is an aim of the present invention to provide an improved power amplifier to overcome some of the above problems.
[0006] According to a first aspect of the invention, an amplifier is provided comprising power selection means for selecting a nominal output power level from a plurality of discrete levels; power control means for controlling output power in response to said power selection means; input means for indicating measured output power; and storage means responsive to the input means for storing, for each of said nominal output power levels, information in response to the measured output power, for future adjustment of each nominal output power level when selected.
[0007] According to a preferred feature of the invention, the amplifier comprises: power selection means for selecting a nominal output power level from a plurality of discrete levels; power control means for controlling output power in response to said power selecting means; input means for indicating measured output power; and storage means responsive to the input means for storing information in response to the measured output power, for future adjustment of the selected nominal output power level. The information stored may be an indication, in respect of each of said nominal output power levels, as to which of a plurality of sub-levels, offset from said selected level, gives rise to an output power closest to that nominal output power level.
[0008] In this manner, whichever of the sub-levels best representing the desired output power level is selected. As an alternative to providing preprogrammed sub-levels, preprogrammed or dynamic offsets can be used, which are added to the nominal power level values. No manual adjustment is required. The storage means records which of the sub-levels (or what offset) is to be used and that sub-level (or offset) is used thereafter. The remaining sub-levels remain unused. This facilitates calibrating of the power levels before the equipment leaves the factory. It also makes recalibration of the equipment quick and simple. With modification, recalibration could be carried out automatically by the equipment itself. It also allows for dynamic power control by changing from one sub-level to another (or by changing the offset) during use to compensate for drift, temperature etc. The storage means may record, from one time slot to another, an indication of the measured output power so as to control the output power in a later time-slot.
[0009] The amplifier may be a pulsed power amplifier which is responsive to a control pulse for amplifying a radio frequency signal comprising: means for producing a predetermined sequence of values in response to each control pulse; means for converting each value to a power control signal; and control means for controlling the amplifier output power in accordance with said power control signal to provide a predetermined amplifier response function.
[0010] By this means, the time domain template on turn-on and turn-off is governed by the sequence of values. Any desired power/time characteristics can be selected and, in principle, the accuracy of the characteristic is dictated only by the number of values used and the accuracy of the samples. The Applicant has found that a raised cosine shape for the RF voltage gives rise to the minimum spectral noise. Accordingly, the samples can be determined so as to give rise to this characteristic, taking into account any non-linearities in translating the values into a RF voltage. As an alternative to a raised cosine, a gaussian shape can be used.
[0012] The above features can conveniently be implemented using a single shaping ROM. For example, for sixteen levels, each having four sub-levels, the ROM merely has to store sixty-four power/time characteristics.
[0013] Preferably a feedback control loop is provided comprising sensing means for sensing output power and comparator means for receiving and comparing an output power signal from said sensing means and an output power level determining signal, wherein said power control means are arranged to control the output power so as to equalise said signals. Whereas a digital comparison (subtraction) could be made, it is preferred that said comparator means are arranged to receive said output power signal on a first input and said power determining signal on a second input, said inputs being connected to a common voltage level by means of two diodes, said diodes being adjacent each other in substantially isothermal relationship. In this manner, variations in thermal characteristics of the diode detector are effectively cancelled out. In the preferred embodiment, the output power level determining signal is derived via a digital-to-analog converter from the shaping ROM, and the feedback signal is derived from the output of the power amplifier.
[0014] A preferred embodiment of the invention will now be described with reference to the accompanying drawings, in which:
Figure 1 shows a power amplifier for a radio transmitter, in accordance with the present invention;
Figure 2 is a block diagram showing the power amplifier of Figure 1 during calibration of power levels;
Figure 3 shows a typical desired signal on feedback loop 13.
Figure 4 shows a circuit for use in an alternative embodiment of the invention; and
Figures 5 and 6 show further circuits for use in alternative embodiments of the invention, incorporating a variable time base.
[0015] Referring to Figure 1, an RF section 1 is shown and a power control section 2. The RF section has an input 3 for receiving data to be transmitted and an output 4 for providing an RF signal for transmission. The RF signal is fed to attenuator 9 and RF power amplifier 10. The output of power amplifier 10 is fed to the antenna 11. From the output of the power amplifier 10, there is also a level sensor 12, which is connected to a feedback loop 13 in the power control section 2.
[0016] The power control section 2 has a six-bit power control input 15, which is connected to the address lines of a shaping ROM 16. The power control section 2 also has a clock input 17, which is fed to a six-bit counter 18 which in turn is connected to a further six address bits of the shaping ROM 16. A ramp control unit 19 is connected to the input of the six-bit counter 18 and is controlled by the clock 17 and a transmit input 20. ROM 16 provides an eight-bit output which is fed to a digital-to-analog converter 21, from which the resulting analog signal is fed via comparator amplifier 22 to the attenuator 9 of the power amplifier section 1. The negative input of comparator amplifier 22 is connected to the level sensor 12 via the feedback loop 13. Each of the inputs of the comparator amplifier 22 has a biasing diode 23a and 23b, connecting it to ground. The diodes 23a and 23b are in close thermal contact on the same chip. This feature has the advantage of eliminating the thermal coefficient of the diode detector. A transmit-sense line 24 is provided, leading from the output power level sensor 12, via a level detector 25 to the transmit controller 30 (Figure 2)
[0018] The transmitter transmits at a frequency from 890-915 MHz and receives at a frequency 45 MHz higher. The transmitter is active for approximately one time slot in every frame. A frame is 4.615 ms long and consists of eight time slots. The time slot duration is 577 us, which is 156.25 bits. The transmitter is active for only 147 bits or 543 us.
[0019] To transmit, the transmit controller 30 selects a power level on power level control lines 15, provides a transmit control pulse on line 20 and provides data to be transmitted on input 3. The output power template, i.e. the output power/time characteristic, is controlled by counter 18, ramp control 19 and shaping ROM 16. When the transmit key 20 indicates start of transmission (S - Figure 3), ramp control 19 controls start of ramping, whereupon it counts 64 pulses (or some other number) to terminate ramping. When the transmit key indicates end of transmission (E), ramp control 19 causes counter 18 to ramp down again through a different set of values. As the transmit pulse progresses, counter 18 counts the input clock pulses 17 and addresses ROM 16 accordingly. Thus, for a given nominal power level, ROM 16 dictates the output characteristic and the output power is controlled accordingly by means of digital-to-analog converter 21, comparator 22 and attenuator 9. For a different nominal power level, a different characteristic is addressed by means of different addresses on power control input 15.
[0020] The six power level control bits at input 15 serve to ease the achievement of correct output power levels. There are sixteen nominal power levels and each nominal power level is split into four sub-levels close to the nominal value. Periodically, the transmit controller carries out an output power test, during which it sets the power to all the 64 possible power output values in turn. The corresponding output powers are measured by external power measuring means in the form of calibrating apparatus 31. The transmit controller is then told by means of an input 32 which of the sub-levels is the best to represent each one of the sixteen nominal output power levels. The result is stored in storage means in the transmit controller 30. Thereafter, the actual output power levels will be correct.
[0021] The characteristic stored in the shaping ROM is an approximation to a raised cosine. By this means, the power up/down ramp is slowed down, in order to reduce the spectral noise in adjacent channels due to the burst modulation. The degree of approximation to the cosine is limited by the step nature of the characteristic stored in the ROM 16.
[0022] The above description has been given by way of example only, and modification of detail can be made within the scope of the invention. Thus, for instance, the power templates stored in ROM 16 could be sub-divided into fewer or more time divisions by decreasing or increasing the clock rate 17 and selecting the count ratio of counter 18 accordingly. Likewise, fewer or more power sub-levels could be provided, and the number of power level control lines 15 and capacity of ROM 16 would need to be selected accordingly. Likewise, greater or lesser accuracy can be achieved from ROM 16 by providing more than eight bits or less than eight bits to the digital-to-analog converter 21.
[0023] The above features of sampling rate variation and resolution could be adapted to, or made a function of, different power levels or other parameters.
[0024] The power amplifier is not solely applicable to QPSK transmitters, nor even to burst modulated transmission. The amplifier could be used in radio transmitters other than for the GSM network, for example in two-way radio. Thus, for power level control of a continuous signal, counter 18 and ramp control 19 can be omitted, leaving a much reduced ROM 16, which merely stores the power levels for the four sub-levels of each of the sixteen nominal power levels. Similarly, for control of a burst modulated transmission at a single power level, power level control lines 15 could be omitted.
[0026] To avoid generating step noise and glitches potentially arising from digital steps in power level, a simple integrator can be used to convert a step input into a slope that is linear with respect to time. Usually, however when an integrating amplifier is operating at a supply rail, it is slow in responding, and also the negative input is not at virtual ground, enabling some coupling of the input to the output. Figure 4A shows the use of a pair of back-to-back zener diodes, Z1 and Z2, that will limit the output to plus or minus the zener voltage, and keep the input at virtual ground. This circuit generates ramps that are determined solely by R1 and C1 and the input amplitude.
[0027] Figure 4B shows a circuit in which the effective value of R1 is modulated (by selectively switching R2-R5 into parallel connection with R1) and C1 and the input amplitude are held constant. The input signal is derived from a CMOS gate of negligible resistance (compared to R1), and thus of constant amplitude (+6 to ground). The positive input of the operational amplifier 40 is biassed to half of the CMOS voltage, so that the input swing relative to the virtual ground is symmetrical. The output will swing from this reference up approximately Z1 volts and down approximately Z2 volts, (plus a little more due to forward diode drops). For the purposes of describing the operation, the Key signal enters at a 74HC04, which, from a logic input, produces a step from +6 volts Off to ground On and back to +6 volts at turn-off. R1-C1 develops a very gentle ramp, so that just before a step is to be executed the output will be on a rail. R2-R5 are all lower value resistors than R1, in the ratio 8:4:2:1, so that in combination of one or more, will develop fifteen different net values of resistance against which C1 can work to develop ramps of different slopes, and are switched so as to modulate the slope of the output waveform.
[0028] There are many ways to generate the slope switching. For purposes of explanation, a programmable array logic (PAL) for a common table look-up and count control is employed. An oscillator provides a clock fast enough to provide a multiple of pulses to an up/down counter during a ramp. It will advance the counter until the table look-up reaches a prescribed count, at which point the table cuts off further counting until key-down is sensed, at which time the counter will count down. The counter's state is combined with the key signal in the PAL to provide a translation to slope, so that the slope profile can be different for key-up and key-down, and need not dwell equally on each slope increment, or indeed even use all of the 15 increments available in this embodiment. Indeed, it may even be desirable to use more than four switched resistors (of binade ratio) or use some other ratio.
[0029] The PAL also provides a test override so that during testing, external signals have control of the slope. These are arranged so that if no external signals are connected when the test input is grounded, the slope will be maximum. Slope maximum is useful in determining the proper value for C1.
[0030] The embodiment of Figure 4 is capable of generating a smoother transition with fewer steps than the embodiment of Figure 1.
[0031] Figure 5 illustrates a further embodiment of the invention. In this embodiment, a high rate digital clock 50 feeds a variable modular counter 51, which, when keyed down, divides by 1 or 2, thus providing a high rate clock having selectable clock rates to a binary ramp counter 52. The counter 52 is locked from counting until key-up (point S in Figure 3). The counter feeds a digital-to-analog converter 53, the filtered output of which controls the RF power level. The D/A converter also feeds a modulo translation table 54, which establishes how many digital clocks are required to advance the binary ramp counter 52 by one step. A controlling microcomputer 55 loads the modulo translation table 54 with the desired ramp up and down information for all the steps, including key-up transmit time and key-down. Upon a start command (to key up the transmitter), the counter 52 steps off. The period of each step thereafter becomes a function of the translation.
[0032] As a practical matter, the digital clock 50 must be faster that the desired ramp speed. A 50 MHZ clock could usually provide about 100:1 time base to a ramp in the 10-50 microsecond range. An alternative method would use a VCO 60 as depicted in Figure 6, controlled by a linear D/A converter 61 driven from the translation table 64. The range of the VCO might be expanded by mixing and offsetting it. For example a VCO spanning the range 50-60 MHZ mixed against a 49 MHZ signal will yield 1-11 MHZ, more linearly than could easily be generated from a 1-11 MHZ VCO directly.
[0033] Instead of controlling attenuator 9 with the signal from comparator 22, a power amplifier with variable gain control can be used and the signal from comparator 22 can adjust the gain.
1. An amplifier comprising:
power selection means (15) for selecting a nominal output power level from a plurality of discrete levels;
power control means (16, 21, 22, 9, 12, 13) for controlling output power in response to said power selection means;
input means (32) for indicating measured output power; characterized by
storage means (30) responsive to the input means for storing, for each of said discrete levels, information in response to the measured output power, for future adjustment of each nominal output power level when selected.
power selection means (15) for selecting a nominal output power level from a plurality of discrete levels;
power control means (16, 21, 22, 9, 12, 13) for controlling output power in response to said power selection means;
input means (32) for indicating measured output power; characterized by
storage means (30) responsive to the input means for storing, for each of said discrete levels, information in response to the measured output power, for future adjustment of each nominal output power level when selected.
2. An amplifier according to claim 1, wherein the power selection means comprises means
for selecting power sub-levels offset from said selected nominal output power level;
and wherein the storage means comprises means for recording, in respect of each of said nominal output power levels, which of said sub-levels gives rise to an output power closest to that nominal output power level.
and wherein the storage means comprises means for recording, in respect of each of said nominal output power levels, which of said sub-levels gives rise to an output power closest to that nominal output power level.
3. An amplifier according to claim 2, further comprising a transmitter controller (30)
arranged to select a nominal transmit power level, select various power sub-levels
offset from that level, store information indicative of the said closest sub-level
and repeat the process for each other nominal transmit power level in turn.
4. An amplifier according to claim 2, wherein each power level and sub-level is represented
as a sequence of values (19).
5. A power amplifier according to any one of the preceding claims, wherein a feedback
control loop (22, 12, 13) is provided comprising sensing means for sensing output
power and comparator means for receiving and comparing an output power signal from
said sensing means and an output power level determining signal, wherein said power
control means are arranged to control the output power so as to equalise said signals.
6. A power amplifier according to claim 5, wherein said comparator means (22) are arranged
to receive said output power signal on a first input and said power determining signal
on a second input, said inputs being connected to a common voltage level by means
of two diodes (23A, 23B), said diodes being adjacent each other in substantially isothermal
relationship.
7. A method of controlling output power in amplification of a radio frequency signal
by a pulsed power amplifier, comprising the steps of:
selecting (15) a nominal output power level from a plurality of discrete levels;
controlling output power in response to said selection of level and sub-levels; and
measuring (31) the actual output power, storing (30) information in response to the measured output power for each of the plurality of discrete levels and adjusting the nominal output power level at a later time dependent on the stored information.
selecting (15) a nominal output power level from a plurality of discrete levels;
controlling output power in response to said selection of level and sub-levels; and
measuring (31) the actual output power, storing (30) information in response to the measured output power for each of the plurality of discrete levels and adjusting the nominal output power level at a later time dependent on the stored information.